Azati Team, a participant in the NASA International Space Apps Challenge, was chosen to receive the Local People’s Choice Award and to become a Global Award Nominee finalist. The NASA International Space Apps Challenge is the world’s largest hackathon that took place on April 29 through April 30, 2017. Nearly 25,000 participants around the world had 48 hours to create open-source apps addressing the challenges posed by NASA.
The Azati team had selected the “You are my Sunshine” challenge, aiming to create a medium to help people understand energy output from a solar panel, and a tool to plan energy consumption based on expected energy output from solar technologies. The project, named “Itibo SolarCity,” used advanced artificial intelligence (AI) methodologies and applied neural networking technique, tailored to process satellite images. The immediate goal of the project is to automatically identify roofs suitable for solar panel installation within a selected neighborhood, town or city.
The application’s self-learning capabilities were used to define insolation incident on surface data available from NASA Global Data Sets. The Itibo SolarCity project team defined the necessary parameters needed for estimating the effective roof surface area, making them suitable for solar panel installations. Suitable roofs detection is performed by a neural network using transfer learning techniques. A set of features is extracted from roofs’ images by a deep convolutional pre-trained 50-layer neural network ResNet-50. The resulting features are then processed using support vector machine algorithm, thus facilitating successful extraction and identification of the following information:
- number of roofs,
- roof total surface area,
- solar panel total surface area,
- energy production by month and average energy production per roof.
The Azati team, by creating the Itibo SolarCity project, aims to provide information needed for creating investment and/or subsidizing strategies, thus facilitating the industry and government planning of future energy systems. The project can be further developed by considering other unused space within a town (e.g. poles and walls) and scaling the calculations within regions and countries. Additionally, beyond the Itibo SolarCity project, similar artificial intelligence (AI) neural networking technologies are also applied to serve commercial clients in need of processing large datasets.
Azati is an IT consulting and software development company based in the United States and Eastern Europe. Azati serves clients in multiple industries, including advertising, marketing, biotechnology, financial, insurance, and other fields. Azati stays on the cusp of advanced information technology trends, including artificial intelligence and neural networking, but without abandoning the more traditional needs of its clients, such as systems integration, web and mobile development.